The FFQ is a reliable and valid instrument for collecting alcohol intake in regular drinkers. Lower validity in irregular drinkers may be due to seasonal variation and/or inadequacy of the FFQ to capture irregular patterns of consumption and/or inadequacy of the average of two 7-DD as a reference method. Furthermore, a considerable degree of misclassification was observed between non-drinkers and moderate drinkers.

Histograms of all age-standardized (world population) death certification rates for 23 cancers or groups of cancers for the period 1990-92 were produced for 35 countries of the European region (including a dozen new national entities) providing data to the World ealth Organization database. Substantial variations were observed in mortality from most common sites. For lung cancer the rate in males was 81/100,000 in Hungary, followed by Belgium, the Czech Republic, the Russian Federation and Poland, while in Sweden, Iceland and Norway, where comprehensive antismoking campaigns have been adopted over the last two decades, the rates were between 24 and 30 per 100,000 males. The lung cancer epidemic in European females is still in its early phases in most countries, with the sole exception of Scotland (29/100,000, ie the highest rates in the world), the rest of the UK, Denmark, Iceland, Ireland and Hungary. With reference to colorectal cancer, the highest rates were in the Czech Republic (38/100,000 males, 21/100,000 females) and other central European countries, and the lowest in Greece, Romania and a few Republics of the former Soviet Union, as well as Finland and Sweden. The highest gastric cancer mortality rates were in the Russian Federation (41/100,000 males, 18/100,000 females), followed by a few Republics of the former Soviet Union and Portugal in Western Europe. The highest breast cancer rates (25-29 per 100,000 females) were in the UK, Belgium, Ireland, The Netherlands, Denmark and other Scandinavian countries. For overall cancer mortality, the range of variation was between 260/100,000 in Hungary and 132/100,000 in Sweden for males, and between 142/100,000 in Denmark and 76/100,000 in Kyrgizstan for females, ie approximately a twofold variation in both sexes.

Data from the Vaud Cancer Registry, Switzerland, were used to analyse incidence and mortality from cancer in teenagers (aged 10 to 19 years) over the period 1974-1992. A total of 113 males and 87 females were registered. Of these, 23% were lymphomas, 16% leukaemias, about 15% central nervous system neoplasms, 10% germ cell tumours and bone neoplasms and 8% soft tissue sarcomas. The overall incidence rate (age-adjusted, world standard population) for all cancers combined was 167 per million boys and 128 per million girls. In both sexes, there was some indication of rising trends over time, to reach 196 per million males and 141 per million females in 1986-1992. The only types of cancer showing consistent upward trends in both sexes were lymphomas. A total of 53 cancer deaths were certified, due to leukaemias in about 40% of cases in both sexes and to lymphomas and brain tumours in 15%-20%. The overall mortality rate at age 10-19 years (age-standardised, world standard) was 47 per million boys and 31 per million girls, and no trend in mortality was observed over time. The 2 main findings of our analysis are (i) the absence of major trends in cancer incidence in adolescents, with the sole exception of a possible increasing incidence of lymphomas, and (ii) the lack of appreciable trends in mortality, in view of the declines in cancer mortality observed in children and young adults over the same calendar period.

Using data from a case-control study conducted between 1985 and 1992 in northern Italy on 828 cases of colon cancer, 498 cases of rectal cancer and 2,024 controls in hospital for acute, non-neoplastic, non-digestive tract disorders, we estimated the percent population attributable risk (PAR) for colorectal cancer in relation to beta-carotene, vitamin C (as markers of a diet rich in fruit and vegetables), red meat and seasoning fat intake, daily meal frequency and family history of the disease. On the basis of multivariate odds ratios, adjusted for total calorie intake, a low intake of beta-carotene accounted for 39% of all the cases and a low intake of vitamin C for 14%. These two micronutrients together explained 43% of all colorectal cancer cases in this population. A high frequency of intake of red meat consumption explained 17% of all cases, and a high score of seasoning fats 4%. A higher daily meal frequency was responsible for 13% of the cases, and these 5 dietary factors together explained 63% of colorectal cancer cases in this population. Family history of colorectal cancer accounted for 4% of all cases. These estimates were similar for colon and rectal cancers separately, in males and females, and in younger and elderly subjects, except for seasoning fats and family history, whose PARs were apparently greater for colon cancer and at younger age. Thus, even though available dietary data were limited in several aspects, and the PAR estimates were based on somewhat arbitrary assumptions regarding the exposure distribution, about two-thirds of all colorectal cancers in this population could be explained in terms of a few risk factors or risk indicators considered. This would correspond to the avoidance of a large proportion of the over 18,000 deaths from colorectal cancer registered per year in the whole of Italy.

Smoking prevalence and patterns in Italy were analyzed using data from the 1990-1991 Italian National Health Survey, based on a sample of 27, 135 males and 28,854 females aged 15 years or over, randomly selected within strata of geographic area and size of the place of residence and of the household, in order to be representative of the general Italian population. Overall, 26.9% of the Italians aged 15 years or over described themselves as current smokers (37.2% males, 17.4% females), and 14.0% as ex-smokers (22.2% males, 6.4% females). The difference in smoking prevalence between males and females was 65% below age 45, but increased substantially with increasing age up to 5-fold above age 65. Moderate smokers (< 15 cigarettes per day) were 12.6% of males and 10.4% of females, intermediate smokers (15 to 24 cigarettes per day) 17.7% of males and 5.5% of females, and heavy smokers (> 25 cigarettes per day) 6.3% of males and 1.5% of females. Pipe or cigar smokers were 0.6% of males. The averange number of cigarettes per smoker per day was 16.6 (17.9 for males, 14.0 for females). The overall smoking prevalence of 26.9% was the lowest registered since 1949, thus confirming the long-term steady decline of smoking, particularly among males. Smoking prevalence, however, has remained constant over the last 15 years among females, after substantial rises in previous calendar years. These falls in overall self-reported smoking prevalence were reflected in declines of legal sale figures (-15% between 1986 and 1991), although it is difficult to quantify the impact of smuggling on total tobacco consumption. Thus, at least part of the falls in self-reported tobacco consumption is attributable to increased underreporting. In males, but not in females, smoking was less frequent in northern (and wealthier) areas of the country, and in more educated individuals. The opposite pattern was observed in females, indicating that even more educated Italian women have not yet recognized the accumulated evidence on the health consequences of smoking. These patterns in smoking are reflected by recent trends in lung cancer, which show some decline in males but persistent upward trends in females, although still on much lower absolute values.

Using data from a case-control study conducted between 1986 and 1992 in Northern Italy on 367 cases of laryngeal cancers (350 men and 17 women) and 1931 hospital controls (1373 men and 558 women), we estimated the relative risks (RR) and the population attributable risks (AR) for laryngeal cancer in relation to tobacco and alcohol consumption and a diet containing little fresh fruit and vegetables (low beta-carotene intake). In men, the RR and their 95% confidence interval (CI), derived from multiple logistic regression, including terms for center, age, and education, plus, simultaneously, tobacco, alcohol, and beta-carotene) were 3.3 (95% CI, 1.9 to 5.5) for ex- or moderate smokers (< 15 cigarettes/day) and 8.8 (95% CI, 5.2 to 14.8) for heavy current smokers compared to never smokers; the RR were 1.5 (95% CI, 1.0 to 2.2) for drinkers of 6 to < 8 alcoholic drinks/day and 2.2 (95% CI, 1.6 to 3.0) for drinkers of 8 or more drinks/day compared to teetotallers or moderate drinkers; with respect to carotene intake the RR were 1.4 (95% CI, 1.0 to 2.0) for the intermediate tertile and 1.8 (95% CI, 1.3 to 2.5) for the lowest tertile, as compared to the highest tertile of intake. AR were estimated assuming a multiplicative model: the single factor with the largest AR was smoking, which accounted for about 77% of laryngeal cancers in men; alcohol intake explained about 25% of cases, low beta-carotene intake accounted for about 18% of cases, and together the three factors were responsible for about 86% of laryngeal cancers in men.(ABSTRACT TRUNCATED AT 250 WORDS)

The relation between breast feeding and breast cancer was investigated in a multicentric case-control study conducted in Italy on 2,167 parous women with histologically confirmed breast cancer, diagnosed within 1 year, and 2,208 parous control women admitted to hospitals in the same catchment areas of cases for acute, non-neoplastic, non-gynecological non-hormone-related diseases. Compared with women who had never tried to lactate, those who had always failed had a multivariate odds ratio (OR; adjusted for parity, education and several other potential confounding factors) of 0.94, and those who had lactated had an OR of 1.17. The multivariate ORs of women who had breast fed 1, 2 and 3 or more children were, respectively, 1.14, 1.18 and 1.32, compared with women who had never lactated. None of these ORs was statistically significant. Compared with women who had never breast fed, the multivariate ORs were 1.19 for women reporting less than 6 months of breast feeding, 1.15 for 6–11 months, 1.34 for 12–17 months, 1.10 for 18–23 months and 0.86 for 24 months or more. No appreciable difference was evident across strata of age, menopausal status, parity and age at first birth, while there was a hint of interaction with education. Our study therefore excluded any appreciable protective role for lactation in breast cancer risk, with the patterns of lactation in this European population, aside from the protective role of parity on breast carcinogenesis. © 1996 Wiley-Liss, Inc.

These findings provide evidence of a satisfactory validity of self-reported smoking habits in pregnancy.

The role of constitutional and environmental factors on the risk of non-melanomatous skin cancer was evaluated in a case-control study conducted in 1992 in Alexandria, Egypt, on 136 incident histologically confirmed (99 basal-cell and 37 squamous-cell) cases of non-melanomatous skin cancer (NMSC) and 145 controls in hospital for a broad spectrum of acute non-sun-related dermatological conditions. In relation to skin colour, compared with brown-skinned subjects, the multivariate relative risks (RR) were 2.3 for olive-skinned subjects and 3.8 for fair/medium-skinned subjects. Three cases and 29 controls were black (RR = 0.2). The trend in risk with skin colour was significant. Likewise, compared with subjects with brown or hazel eyes, those with green or blue eyes had a RR of 3.1. In relation to acute sun reaction, compared with subjects reporting easy tanning, the RRs were 2.5 for subjects reporting moderate tanning and 4.7 for those reporting easy burning. The risk of NMSC was higher for subjects reporting an outdoor occupation than for those reporting an indoor occupation (RR = 7.7). A significant trend in risk was observed with degree of sun exposure: compared with subjects reporting light sun exposure, the RR was 3.0 for those reporting moderate exposure, and 6.1 for those reporting heavy sun exposure. There was an indication of a relationship between clothing pattern and skin-cancer risk: compared with subjects reporting frequent use of traditional Egyptian clothes, the RR for dressing in short clothes was 1.8. The presence of signs of photodamage was also associated with NMSC (RR = 3.7). Exposure to arsenic was reported by 10 cases and 1 control (RR = 9.5). A positive interaction between sun exposure and skin colour was observed, and the RR rose to 14.2 for medium- or fair-skinned subjects with heavy exposure compared with brown- or black-skinned subjects with light or moderate sun exposure. In this Egyptian population, over 60% of NMSC could be attributed to sun exposure and approximately 45% to skin colour.

After early rises, oesophageal cancer mortality rates have tended to level off in Italy over the last two decades, and are now intermediate on a European scale (4.7/100000 males, 0.8/100000 females, world standard). This reflects the trends in consumption of the major risk factors for the disease, tobacco and alcohol, particularly in men. Within Italy, there is also a substantial variation in oesophageal cancer rates, with high mortality areas in the North-East of the country. On the basis of a case-control study conducted in northern Italy, the relative risk (RR) of the disease was 4.3 in heavy smokers and 3.5 in heavy drinkers. A diet poor in fresh fruit and vegetables was also related to the risk of the disease (RR = 2.5). In terms of population attributable risk, 71% of cases in men and 32% in women were accounted for tobacco smoking, 45% in males and 10% in females by alcohol drinking, 40% in men and 29% in women by a diet poor in fresh fruit and vegetables (and hence in beta-carotene). These three factors together accounted for 90% of cases in Italian men, and 58% in women (83% in both sexes combined).

Since gastrectomy may cause gastric hypoacidity and favour oesophageal reflux, and hence oesophagitis, the results of this study, if confirmed, are of interest to a better understanding of the mechanisms of oesophageal carcinogenesis.